property modification factors for seismic isolators: design guidance for buildings
TRANSCRIPT
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
1/242
SSN 1520-295X
Property Modication Factorsfor Seismic Isolators:
Design Guidance for Buildings
by
William J. McVitty and Michael C. Constantinou
Technical Report MCEER-15-0005
June 30, 2015
This research was conducted at the University at Buffalo, State University of New York, and was
supported by MCEER Thrust Area 3, Innovative Technologies.
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
2/242
NOTICEThis report was prepared by the University at Buffalo, State University of NewYork, as a result of research sponsored by MCEER. Neither MCEER, associatesof MCEER, its sponsors, the University at Buffalo, State University of NewYork, nor any person acting on their behalf:
a. makes any warranty, express or implied, with respect to the use of anyinformation, apparatus, method, or process disclosed in this report or thatsuch use may not infringe upon privately owned rights; or
b. assumes any liabilities of whatsoever kind with respect to the use of, or the
damage resulting from the use of, any information, apparatus, method, orprocess disclosed in this report.
Any opinions, findings, and conclusions or recommendations expressed in thispublication are those of the author(s) and do not necessarily reflect the viewsof MCEER or other sponsors.
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
3/242
Property Modification Factors for Seismic Isolators:Design Guidance for Buildings
by
William J. McVitty1and Michael C. Constantinou2
Publication Date: June 30, 2015
Submittal Date: March 20, 2015
Technical Report MCEER-15-0005
MCEER Thrust Area 3: Innovative Technologies
1 Structural Design Engineer, KPFF Consulting Engineers, Seattle, WA: former graduate student,Department of Civil, Structural and Environmental Engineering, University at Buffalo, StateUniversity of New York
2 SUNY Distinguished Professor, Department of Civil, Structural and Environmental Engineer-ing, University at Buffalo, State University of New York
MCEERUniversity at Buffalo, State University of New York212 Ketter Hall, Buffalo, NY 14260E-mail: [email protected]; Website: http://mceer.buffalo.edu
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
4/242
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
5/242
iii
PREFACE
MCEER is a national center of excellence dedicated to the discovery and development of new knowl-edge, tools and technologies that equip communities to become more disaster resilient in the face ofearthquakes and other extreme events. MCEER accomplishes this through a system of multidisciplinary,
multi-hazard research, in tandem with complimentary education and outreach initiatives.
Headquartered at the University at Buffalo, The State University of New York, MCEER was originallyestablished by the National Science Foundation in 1986, as the first National Center for EarthquakeEngineering Research (NCEER). In 1998, it became known as the Multidisciplinary Center for Earth-
quake Engineering Research (MCEER), from which the current name, MCEER, evolved.
Comprising a consortium of researchers and industry partners from numerous disciplines and institutionsthroughout the United States, MCEERs mission has expanded from its original focus on earthquakeengineering to one which addresses the technical and socio-economic impacts of a variety of hazards,both natural and man-made, on critical infrastructure, facilities, and society.
The Center derives support from several Federal agencies, including the National Science Foundation,
Federal Highway Administration, National Institute of Standards and Technology, Department ofHomeland Security/Federal Emergency Management Agency, and the State of New York, other stategovernments, academic institutions, foreign governments and private industry.
This report provides guidance on the application of the provisions of the 2016 ASCE 7 Standard for the analysis anddesign of seismically isolated buildings. These new provisions introduced a systematic approach for the determination ofthe bounding values of the mechanical properties of the isolators on the basis of experimental data of prototype isola-tors and considerations for the effects of the environment, aging and uncertainty. The report presents an overview of the
concept of system property modification factors that is used in the establishment of the bounding values and proceeds withtwo systematic examples, one for an elastomeric and one for a sliding isolation system. Moreover, the new provisions arecritically reviewed for consistency between the requirements for establishing bounding values and the prescribed acceptancecriteria in the testing of prototype isolators. Changes are proposed to the standard to avoid inconsistencies.
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
6/242
iv
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
7/242
v
ABSTRACT
The application of seismic isolation in the United States is regulated by building codes which invariably
refer to the ASCE 7 Standard for analysis and design requirements. The Standard has evolved over the
years to reflect the state of the art and practice in the field. The latest evolution, ASCE 7-2016, now
includes a systematic procedure for establishing upper and lower bound values of isolator properties with
due consideration for three categories of effects, namely: a) aging effects and environmental conditions,
b) hysteretic heating and history of loading effects, and c) manufacturing variations.
This report provides commentary to practicing engineers on the basis and implementation of ASCE 7-
2016 (the Standard), with respect to the property modification (lambda or ) factors. These factors are
used to define the upper and lower bound properties for analysis and design. It is emphasized, however,
that this report does not strictly follow all provisions of ASCE 7-2016. Rather, it includes an evaluation
and a critique of the provisions on the procedures of obtaining data on the bounding values of mechanical
properties, and recommendations for changes. The report starts with discussion on the mechanical
properties of isolators, their force-displacement models and the concept of property modification factors.
Next there is clarification on the types of property modifications mentioned in the Standard with guidance
on the magnitude of various effects. An interpretation of the Standards testing requirements is presented,
with direction given on how to decide on the lower bound values of mechanical properties for analysis.
The determination of property modification factors is illustrated for a elastomeric and a sliding isolation
system which are made up of lead-rubber and low-damping natural rubber isolators and triple Friction
PendulumTM isolators, respectively. For each system, the following design scenarios are reported: (a)
assuming there is no qualification test data available, and (b) using prototype test data of two isolators. A
third option (c) of having complete production test data available for the analysis and design is also
discussed. These factors are project-specific, manufacturer-specific and also dependent on the material
used, therefore cannot be simply adopted for other designs. Finally, nonlinear response history analyses
are undertaken for each design scenario (i.e. each range of isolation system properties) to show the effects
of bounding.
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
8/242
vi
This Page is Intentionally Left Blank
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
9/242
vii
ACKNOWLEDGEMENTS
This report received a thorough examination from Martin Button, Ph.D., PE, a consulting engineer with
Button Engineering of Austin, Texas. The authors gratefully acknowledge his comments and great
contribution to improving the quality and clarity of the report. Martin Button was a member of the ASCE
7-2016 committee that revised Chapters 17 and 18.
The opportunity for the first author to undertake graduate studies at the University at Buffalo was
provided by the Fulbright Program and the New Zealand Earthquake Commission (EQC), through the
Fulbright-EQC Award in Natural Disaster Research. The authors appreciate the support and recognize the
value of this educational and cultural exchange. It has provided the chance to share knowledge on the
state of practice internationally, and to collaborate to provide guidance which is applicable to engineers
both in the United States and New Zealand.
The participation of the second author in the ASCE 7 revision was sanctioned and partially sponsored by
the Multidisciplinary Center for Earthquake Engineering Research (MCEER).
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
10/242
viii
This Page is Intentionally Left Blank
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
11/242
ix
TABLE OF CONTENTS
SECTION 1 INTRODUCTION................................................................................................... 1
1.1 Scope of Report ................................................................................................................ 1
1.2 Intent of Seismic Isolation Provisions .............................................................................. 2
1.3 Analysis Procedures for Seismically Isolated Structures ................................................. 3
SECTION 2 FORCE-DISPLACEMENT MODELS OF ISOLATORS .................................. 5
2.1 Introduction ...................................................................................................................... 5
2.2 Modeling Elastomeric Isolator Units ............................................................................... 5
2.3 Modeling Sliding Isolator Units ....................................................................................... 7
2.4 Property Modification Factors ....................................................................................... 12
SECTION 3 TYPES OF PROPERTY MODIFICATIONS ................................................... 17
3.1 Introduction .................................................................................................................... 17
3.2 Aging and Environmental Effects: ae,max and ae,min...................................................... 17
3.3 Hysteretic Heating and History of Loading Effects: test,max and test,min........................ 21
3.4 Permissible Manufacturing Variations: spec,max and spec,min.......................................... 38
SECTION 4 TESTING REQUIREMENTS ............................................................................. 43
4.1 Introduction .................................................................................................................... 43
4.2 Qualification Testing ...................................................................................................... 43
4.3 Prototype Testing ........................................................................................................... 44
4.4 Production Testing ......................................................................................................... 47
SECTION 5 PRELIMINARY DESIGN AND DESIGN SCENARIOS ................................. 49
5.1 Introduction .................................................................................................................... 49
5.2 Preliminary Design for Elastomeric Isolation System ................................................... 50
5.3 Preliminary Design for Sliding Isolation System ........................................................... 52
5.4 Design Scenarios for Detailed Analysis ......................................................................... 55
SECTION 6 COMPUTING PROPERTIES FOR ELASTOMERIC ISOLATORS ............ 57
6.1 Introduction .................................................................................................................... 57
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
12/242
x
6.2 Lead-Rubber Isolators .................................................................................................... 57
6.3 Low-Damping Rubber Isolators ..................................................................................... 67
6.4 Similar Unit Criteria ....................................................................................................... 72
6.5 Test Specimen Adequacy Criteria .................................................................................. 77
SECTION 7 COMPUTING PROPERTIES FOR SLIDING ISOLATORS ......................... 83
7.1 Introduction .................................................................................................................... 83
7.2 Prototype Isolators.......................................................................................................... 83
7.3 Similar Unit Criteria ....................................................................................................... 97
7.4 Test Specimen Adequacy Criteria .................................................................................. 99
SECTION 8 SUMMARY OF ISOLATION SYSTEM PROPERTIES ............................... 105
8.1 Elastomeric Isolation System ....................................................................................... 105
8.2 Sliding Isolation System............................................................................................... 113
SECTION 9 NONLINEAR RESPONSE HISTORY ANALYSIS ....................................... 121
9.1 Introduction .................................................................................................................. 121
9.2 Elastomeric Isolation System ....................................................................................... 122
9.3 Sliding Isolation System............................................................................................... 127
SECTION 10 SUMMARY AND CONCLUSIONS ............................................................... 133
SECTION 11 REFERENCES .................................................................................................. 137
APPENDIX A GENERAL BUILDING INFORMATION ................................................... 139
A.1 General Information ..................................................................................................... 139
A.2 Building Geometry ....................................................................................................... 139
A.3 Assembly Weights........................................................................................................ 141
A.4 Design Spectral Accelerations ..................................................................................... 142
A.5 Ground Motions ........................................................................................................... 142
APPENDIX B PRELIMINARY DESIGN OF ELASTOMERIC ISOLATION SYSTEM 145
B.1 Introduction .................................................................................................................. 145
B.2 Isolation Systems Characteristic Strength, Qd,total........................................................ 146
B.3 Isolation System Post-Elastic Stiffness, Kd,total............................................................ 147
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
13/242
xi
B.4 Lower Bound MCE Displacement ............................................................................... 148
B.5 Isolator Axial Loads and Uplift Potential .................................................................... 149
B.6 Rubber Layer Thickness Required for Isolator Stability ............................................. 153
APPENDIX C PRELIMINARY DESIGN OF SLIDING ISOLATION SYSTEM ............ 157
C.1 Introduction ..................................................................................................................... 157
C.2 Isolator Geometric Parameters ........................................................................................ 158
C.3 Design Based on Recommended Isolator Data ............................................................... 159
C.4 Design Based on Qualification Data ................................................................................ 160
APPENDIX D ASCE 7-2016 CHAPTER 17 PROPOSAL WITH CHANGES PROPOSED
BY AUTHORS OF THIS REPORT ....................................................................................... 165
APPENDIX E ASCE 7-2016 COMMENTARY- DEFAULT LAMBDA FACTORS ........ 193
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
14/242
xii
This Page is Intentionally Left Blank
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
15/242
xiii
LIST OF FIGURES
Figure Title Page
2-1 Idealized Bilinear Force-Displacement Relation of a Lead-Rubber Isolator .................................... 6
2-2 Rigid-Linear Force-Displacement Behavior of FP Isolators ............................................................ 8
2-3 Triple Friction Pendulum Cross Section, Definition of Parameters ................................................. 9
2-4 Tri-linear Force-Displacement Behavior of Special Triple FP up to Regime II ............................. 12
2-5 Upper and Lower Bound Isolator Force-Displacement Properties.15
3-1 Concept of Bounding the Hysteretic Heating and History of Loading Effects ............................. 24
3-2 NED for Variaous Isolation System Properties based on Bi-Directional RHA .............................. 26
3-3 Effects of Scragging on a Rubber Isolator ...................................................................................... 30
3-4 Scragging Effects for a Range of Elastomeric Isolator Types ........................................................ 31
3-5 Effect of Hysteretic Heating on the EDC for Lead-Rubber Isolators ............................................. 33
3-6 Effect of Sliding Velocity on the Coefficient of Friction for PTFE-Stainless Steel Interfaces ...... 36
3-7 Effect of Axial Load on the Coefficient of Friction in Friction Pendulum Isolators ...................... 36
3-8 Hypothetical Illustration of Specification Tolerances .................................................................... 40
5-1 Schematic of Isolator Layout for Lead-Rubber Isolation System .................................................. 52
5-2 Schematic of Isolator Layout for Sliding Isolation System ............................................................ 54
5-3 Continuum of Quantity and Quality of Test Data vs. Range in Isolation System Properties ......... 56
6-1 Tested Lead-Rubber Isolator Unit Details ...................................................................................... 58
6-2 Lateral Force-Displacement Loops of Isolator No. 1 ..................................................................... 58
6-3 Lateral Force-Displacement Loops of Isolator No. 2 ..................................................................... 59
6-4 Graphical Calculation of the Post-Elastic Stiffness, Loop 2 Isolator No. 1 ................................... 61
6-5 Measured and Calculated Effective Yield Stress of Lead .............................................................. 65
6-6 Tested Low-Damping Natural Rubber Isolator Details .................................................................. 67
6-7 Lateral Force-Displacement Loops of Isolator No. 3 ..................................................................... 68
6-8 Lateral Force-Displacement Loops of Isolator No. 4 ..................................................................... 68
6-9 Graphical Calculation of the Post-Elastic Stiffness, Loop 1 Isolator No. 3 ................................... 69
6-10 Graphical Calculation of the Post-Elastic Stiffness, Loop 3 Isolator No. 3 ................................... 70
7-1 Cross-Section and Details of Tested Triple Friction Pendulum Isolator ........................................ 84
7-2 Dynamic Force-Displacement Behavior of Prototype Isolator 1: Test 1 ........................................ 86
7-3 Dynamic Force-Displacement Behavior of Prototype Isolator 1: Test 2 ........................................ 86
7-4 Dynamic Force-Displacement Behavior of Prototype Isolator 2: Test 1 ........................................ 87
7-5 Dynamic Force-Displacement Behavior of Prototype Isolator 2: Test 2 ........................................ 87
7-6 Graphical Measure of Properties of Prototype Isolator 1 for Test 1 ...89
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
16/242
xiv
7-7 Graphical Calculation of Properties of Prototype Isolator 1 for Test 2 ............................................ 90
7-8 Estimate of Area for Calculating Friction B, Isolator 1, Test 1, Cycle 1 ........................................ 91
7-9 Theoretical and Actual Force-Displacement Behavior: Prototype Isolator 1, Test 1 ....................... 93
7-10 Three-Cycle Average Coefficient of Friction Aof 43 Production Isolators (1348kN Test 2) ........ 95
7-11 Theoretical and Actual Force-Displacement Behavior for Isolator 2, Test 1 ................................. 101
8-1 Idealized Force-Displacement of Lead-Rubber Isolator using Default Properties ......................... 112
8-2 Idealized Force-Displacement of Lead-Rubber Isolator using Prototype Properties ..................... 112
8-3 Idealized Force-Displacement of Lead-Rubber Isolator using Production Isolator Properties ...... 113
8-4 Tri-linear Model for Interior Isolator using Default Properties ...................................................... 119
8-5 Tri-linear Model for Interior Isolator using Prototype Properties .................................................. 120
8-6 Tri-linear Model for Interior Isolator using Production Properties ................................................ 120
9-1 F-Loops of Elastomeric Isolators for Motion GM3, Default Lambda Factors ........................... 125
9-2 F- Loops of Elastomeric Isolators for Motion GM3, Prototype Lambda Factors ....................... 126
9-3 F- Loops of Elastomeric Isolators for Motion GM3, Production Lambda Factors ..................... 126
9-4 F- Loops of FP Isolation System for Motion GM3, Default Lambda Factors ............................ 130
9-5 F- Loops of FP Isolation System for Motion GM3, Prototype Lambda Factors ........................ 1 31
9-6 F- Loops of FP Isolation System for Motion GM3, Production Lambda Factors ..................... 131
A-1 Typical floor and roof framing plan ............................................................................................. 139
A-2 Three-Dimensional Extruded Sections SAP200 Model ............................................................... 140
A-3 Response Spectrum for Scaled Ground Motions and MCER........................................................ 140
A-4 Response Spectrum for Scaled Ground Motions and MCER........................................................ 143
B-1 Preliminary Design Procedure for Elastomeric Isolation System ................................................. 146
C-1 Special Triple Friction Pendulum Isolator Detail .......................................................................... 158
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
17/242
xv
LIST OF TABLES
Table Title Page
2-1 Force-Displacement Behavior for a Special Triple FPR Isolator................................................. 10
2-2 Schematics of Force-Displacement Behavior for Special Triple FP Isolator............................... 11
3-1 Aging and Environmental Lambda Factors for Elastomeric Isolators ............................................. 19
3-2 Aging and Environmental Lambda Factors for Sliding Isolators .................................................... 21
3-3 Peak Response of Analyzed Structure ............................................................................................. 28
3-4 Testing Lambda Factors for Elastomeric Isolators .......................................................................... 35
3-5 Testing Lambda Factors for Sliding Isolators.................................................................................. 38
3-6 Specification Lambda Factors ......................................................................................................... 41
5-1 Rubber and Natural Rubber Isolator Dimensions obtained from Appendix B ................................ 50
5-2 Elastomeric Isolation System Compression Loads .......................................................................... 52
5-3 Sliding Isolation System Compression Loads ................................................................................. 55
6-1 Mechanical Properties of Two Prototype Lead-Rubber Isolators ................................................... 60
6-2 Mechanical Properties of Two Prototype Low-Damping Rubber Isolators .................................... 71
6-3 Comparison of Governing Dimensions ........................................................................................... 73
6-4 Comparison of Energy Dissipated per Cycle ................................................................................... 74
6-5 Comparison of Shear Strain and Vertical Stress .............................................................................. 76
6-6 Tested Prototype Elastomeric Isolator Shear Moduli Criteria. ........................................................ 79
6-7 Tested Prototype Lead-rubber Isolator No. 1 Properties ................................................................. 81
7-1 Test 1 Data for Prototype Triple FP Isolators .................................................................................. 88
7-2 Test 2 Data for Prototype Triple FP Isolators .................................................................................. 88
7-3 Frictional Properties of Isolator at Load of 2710kN (609kip) from Test 1 ..................................... 93
7-4 Frictional Properties of Isolator at Load of 1348kN (303kip) from Test 2 ..................................... 94
7-5 Effective Stiffness of FP Isolators ................................................................................................. 102
7-6 Effective Damping of FP Isolators ................................................................................................ 103
8-1 Nominal Properties from Prototype Test Data used for all Examples ........................................... 105
8-2 Default Lambda Factors in Absence of Qualification Data for Elastomeric Isolation System ..... 107
8-3 Lambda Factors from Similar Units forElastomeric Isolation System ..................................... 108
8-4 Lambda Factors from Actual Prototype Isolators for Elastomeric Isolation System .................... 109
8-5 Lambda Factors Based on Data from Production Isolators for Elastomeric Isolation System ...... 111
8-6 Nominal Properties from Prototype Test Data used for all Examples ........................................... 114
8-7 Nominal 1and Lambda Factors- Default Values in Absence of Qualification Data ................... 116
8-8 Nominal 1and Lambda Factors- Prototype/Similar Test Data Available .................................... 117
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
18/242
xvi
8-9 Coefficient of Outer Surfaces Friction11and Lambda Factors- Production Test Data ............... 118
9-1 NLRHA of Elastomeric Isolation System using Properties based on Default Factors ............... 123
9-2 NLRHA of Elastomeric Isolation System using Properties based on Prototype Factors ........... 123
9-3 NLRHA of Elastomeric Isolation System using Properties based on Production Factors ......... 124
9-4 Ratio of Lambda Factors and Bounding Analysis for Elastomeric Isolation System .................... 124
9-5 NLRHA of Sliding Isolation System using Properties based on Default Factors ...................... 128
9-6 NLRHA of Sliding Isolation System using Properties based on Prototype Factors ................... 128
9-7 NLRHA of Sliding Isolation System using Properties based on Production Factors ................. 129
9-8 Ratio of Lambda Factors and Bounding Analysis for Sliding Isolation System ........................... 129
A-1 SEAOC Manual Horizontal Ground Motions and Scale Factors ................................................. 143
B-1 Preliminary Design Structural Overturning, Iteration 1 ................................................................ 151
B-2 Preliminary Design Iterations, SI Units ........................................................................................ 155
B-3 Preliminary Design Iterations, Imperial Units .............................................................................. 156
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
19/242
1
SECTION 1
INTRODUCTION
1. 1
Scope of Report
Seismically isolated buildings in the United States are designed in accordance with the requirements of
Chapter 17 of the American Society of Civil Engineers Standard ASCE-7. Currently the 2010 edition of
the Standard (ASCE 7, 2010) is being revised with a number of alterations proposed for Chapter 17 of the
2016 edition (ASCE 7, 2016). Appendices D and E at the end of this report present the revised Chapter 17
during the balloting period that started in late 2014 and was still ongoing at the conclusion of the writing
of this report in February 2015. One major amendment, which is the focus of this report, is the addition
of a new section titled Isolation System Properties. This section consolidates and clarifies previous
provisions and provides new provisions for the explicit development of property modification (lambda or
) factors. These lambda factors are used to determine the upper and lower bound force -displacement
behavior of isolation system components on a project-specific and product-specific basis.
The aim of this report is to provide guidance to practicing engineers on the basis and implementation of
the Standard ASCE 7-2016, with respect to the lambda factors. This is first developed through discussion
on the mechanical properties of isolators, their force-displacement models and how different factors affect
behavior. Next there is clarification of parameters mentioned in ASCE 7-2016 and an interpretation on the
testing requirements, which is complemented with design examples. For each system, the following
design scenarios are investigated: (a) assuming there is no qualification test data available, and (b) using
prototype test data of two isolators. A third option (c) of having complete production test data available
for the analysis and design is also discussed. The lambda factors are calculated for lead-rubber, low-
damping natural rubber and sliding isolators. Finally, nonlinear response history analyses are undertaken
for each design scenario (i.e. each range of isolation system properties) to show the effects of bounding.
The building used in the examples presented in this report is taken from the Structural Engineers
Association of California 2012 Seismic Design Manual, Volume 5 (SEAOC Manual). The SEAOC
Manual is specific to the ASCE 7-10 provisions. This report gives insight on how the latest ASCE 7-2016
provisions may affect the design and gives commentary on meeting the intent of ASCE 7-2016 with
regard to bounding analysis. It is noted that an author of this report was a member of the Standards
committee, and therefore has first-hand accounts of how the ASCE 7-2016 revisions came to fruition.
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
20/242
2
1. 2
Intent of Seismic Isolation Provisions
Seismic isolation is arguably the most effective way to protect a building, its occupants and its contents
from the damaging effects of major earthquakes. The application of the technology in the United States is
regulated by building codes which invariably refer to the ASCE 7 (theStandard)for requirements for
the analysis and design of structures with seismic isolation. The Standard has evolved over the years to
reflect the state of the art and practice in the field. The latest evolution (ASCE 7-2016) includes a
systematic procedure of establishing upper and lower bound values of isolator properties with due
consideration for uncertainties, testing tolerances, aging, environmental effects and history of loading.
The following paragraphs give a perspective of the intent as well as rationale for the Standards
provisions.
The ASCE Standard describes the minimum requirements to provide reasonable assurance of a seismic
performance for buildings and other structures that will:
1. Avoid serious injury and loss of life due to:
a. Structural collapse
b. Failure of nonstructural components or systems
c.
Release of hazardous materials
2. Preserve means of egress
3. Avoid loss of function in critical facilities, and
4.
Reduce structural and nonstructural repair costs where practicable to do so.
These performance objectives do not have the same likelihood of being achieved and depend on a number
of factors. These include structural framing type, building configuration, structural and nonstructural
materials and details, and the overall quality of design and construction. In addition, there are large
uncertainties in the intensity and duration of shaking, which will affect performance.
The key point to note is that seismic isolation is a high-performance system which is not on a level
playing field to other systems. That is, a seismically isolated building and a conventional fixed-base
building, designed to their respective minimum requirements of ASCE 7 will not perform the same in a
major earthquake. A seismically isolated building is expected to have a considerably better performance,
with no significant damage or downtime whereas a conventional building may be damaged to the point
where it is uneconomical to rehabilitate. This is due to the inherent nature of isolation (i.e. reduces both
displacements and accelerations in the superstructure) as well as the minimum requirements of the
Standard. The minimum requirements of ASCE 7-2016 (Chapter 17), although not explicitly seeking
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
21/242
3
damage control as an objective, indirectly will give limited damage as a consequence of ensuring that
proper isolation is achieved. For example, the Standard limits the amount of inelastic action and drift that
may occur in the superstructure in order to avoid detrimental coupling with the isolation system.
Moreover, the Standard requires that the isolators be designed and tested for the effects of the Maximum
Considered Earthquake (MCE).
Seismic isolation can be likened to introducing an engineered soft-story in the structure. All inelastic
action and displacement is concentrated in a single level of the building. Therefore the global response of
the building hinges on the performance of the isolation system, and special care must be taken to ensure it
is reliable. The materials used for isolator units differ somewhat from conventional construction
materials, in that their properties may vary considerably due to temperature, aging, contamination, history
of loading, among other factors (see Section 3). There are no standards which govern how an isolator
must be produced and assembled. These details vary by manufacturer and are usually proprietary.
Furthermore there is no official certification required of manufacturers before they can start supplying
isolators. Consequently, there can be a considerable difference in the quality and performance of isolators,
even for identical isolator types produced by different manufacturers.
Given the importance of the isolation system, and considerable uncertainty in the quality of different
manufacturers, ASCE 7-2016 has taken the approach of requiring the Registered Design Professional
(RDP) to determine (in consultation with the manufacturer) the nominal isolator properties and to account
for the likely variation in those properties on a project-specific and product-specific basis. This is
achieved through a combination of test data and engineering judgment and is now explicitly required in
the Standard through the use of upper bound and lower bound analysis. The intent is that this will give
reasonable assurance of the isolation systems performance throughout the life of the building.
1. 3
Analysis Procedures for Seismically Isolated Structures
There are three methods used to analyze and design seismically isolated buildings. These are the
Equivalent Lateral Force (ELF) method, which is a static procedure, the Response Spectrum method and
Response History Analysis, both of which are dynamic procedures. The practice in the United States has
evolved such that most isolated buildings are designed using Nonlinear Response History Analysis
(NLRHA) with an ELF analysis used to evaluate the results of the dynamic analysis and obtain minima
for response quantities. NLRHA gives the most realistic prediction of response and can be used for all
building types. Conversely, ASCE 7 puts limitations on when the ELF and Response Spectrum methods
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
22/242
4
can be used. Also, ASCE 7-2016 expanded the range of applicability of the ELF procedure relative to
earlier versions of the Standard in an attempt to simplify the application of seismic isolation.
All structural analyses require a mathematical model of the isolators lateral force -displacement behavior.
The adopted nonlinear model can be implemented directly for NLRHA, or it can be linearized using an
effective stiffness and effective damping for the ELF and Response Spectrum procedures. Section 2 gives
common models used for elastomeric and sliding seismic isolators, with detail on which properties govern
the force-displacement behavior. These simple models assume that the mechanical properties of the
isolator remain constant during the earthquake motion as well as during the entire design life. In reality
the properties will vary, perhaps significantly, from the adopted nominal values due to a number of
different effects as discussed in Section 3.
To account for this variation, ASCE 7-2016 takes the approach of performing parallel analyses, one using
the isolation systems upper bound force-displacement properties and one using the lower bound force-
displacement properties. The upper and lower bound values of mechanical properties are determined from
nominal values of properties and the use of property modification (lambda) factors. Even with the most
complex models, which can explicitly account for some property variations (i.e. instantaneous
temperature, velocity, axial load, etc.), there will likely always be a need to perform bounding analysis in
order to account for the effects of aging and specification tolerance, since these factors are not captured
by mathematical models.
The governing case for each response parameter, whether from the upper or lower bound analysis, is then
used for design. Typically the maximum structural forces are from the upper bound analysis and the
maximum isolator displacements are from the lower bound analysis. This approach is similar to capacity-
design, whereby the probable properties of the ductile mechanism are used to design and detail structural
elements. This ensures that (1) a majority of the inelastic deformation and energy dissipation is confined
to the isolation system, and (2) the design can sustain and accommodate the maximum displacement.
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
23/242
5
SECTION 2
FORCE-DISPLACEMENT MODELS OF ISOLATORS
2. 1
Introduction
The analysis of seismically isolated buildings requires a mathematical model of the isolators lateral
force-displacement behavior. The true behavior of an isolator is nonlinear and complex, therefore
different analysis models exist with varying sophistication. This section gives examples of simple models
used to characterize the force-displacement behavior of elastomeric and sliding isolators, and specifically,
what properties/parameters govern the force-displacement behavior. The concept of using property
modification (lambda) factors to envelope the likely isolators behavior is introduced in Section 2.4.
The low-damping natural rubber and lead-rubber elastomeric isolators and the Friction PendulumTM
sliding isolators are the predominant types of seismic isolators used in practice. The desirable features of
these isolators is that laterally they have low stiffness, with good energy dissipation and re-centering
ability, and vertically they remain stable under the weight of the building and large displacements. The
force-displacement behavior is nonlinear, hysteretic and can be idealized by either a rigid-linear, bilinear
or tri-linear model as identified in this section. These types of isolators have been extensively tested and
implemented in practice for the past 30 years, with design based on the models that follow.
It is noted that new isolator concepts cannot be used unless they have verified and validated models. This
involves two parts, (1) the isolators behavior must be verified by dynamic qualification testing in a rig,
and (2) the mathematical model of the isolator used for structural analysis must be validated by
experimental testing (e.g. shake table testing) by showing stability and generation of comparable
analytical and experimental results (i.e. displacements, shear forces, etc.).
2. 2 Modeling Elastomeric Isolator Units
The lateral force-displacement behavior of a lead-rubber isolator can be idealized by the bilinear
hysteretic loop shown in Figure 2-1 (Constantinou et al. 2011).
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
24/242
6
Figure 2-1 Idealized Bilinear Force-Displacement Relation of a Lead-Rubber Isolator
For this basic nonlinear model, the mechanical behavior of the lead-rubber isolator is characterized by the
following three parameters:
a) Characteristic strength, Qd. The characteristic strength is the strength of the isolator at zero
displacement. It is related to the area of leadALand the effective yield stress of lead Las follows:
d L LQ A (2 - 1)
Note that equation (2-1) implies any contribution to the strength from rubber is included in the
effective yield strength of lead. This is a simplification as rubber has a small quantifiable
contribution to strength. The reader is referred to Kalpakidis et al (2008) for a procedure to
account for the contribution of rubber, although the procedure is mostly of academic interest in
advanced modeling of the behavior of lead-rubber isolators.
b) Post-elastic stiffness, Kd. The stiffness is related to the shear modulus of rubber G, the bonded
rubber areaArand the total thickness of rubber layers Tras follows:
rd L
r
GAK f
T (2 - 2)
The parameterfLaccounts for the effect of the lead core on the post-elastic stiffness and ranges in
value from 1.0 to about 1.2. Only after repeated cycling is the value of fLclose to unity. Within
the context of bounding analysis, factorfLis set equal to unity and any uncertainty is incorporated
in the value of G(and by relation in stiffnessKd) by use of the testing lambda factor test.
c) Yield displacement, Y.This parameter dictates the unloading elastic stiffness of the isolator, and
is used in calculating the effective damping. Typically it is in the range of 6 to 25mm (0.25 to 1
inch) and is either assumed or estimated based on a visual fit of test data. Although the in-
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
25/242
7
structure accelerations and the isolator residual displacements can be sensitive to the value of Y, it
is not an important parameter in calculating the global response. Typically only QdandKdare
subject to property modification factors.
The bilinear hysteretic model can be linearized by calculating the effective stiffnessKeff and the effective
damping eff.Keff is calculated by dividing the maximum force by the maximum displacementDMand eff
is calculated as:
2
2 d Meff
eff M
Q D Y
K D
(2 - 3)
Low-damping natural rubber isolators are defined as having an effective damping of less than 5%. The
mechanical behavior of the isolator can be characterized simply as a linear-elastic element with a stiffness
obtained using equation (2 -2) withfLset to unity.A single linear viscous damping element may be added
in each principal direction (and in torsion) to represent the damping capability of all isolators.
Alternatively, each isolator may be modeled as a hysteretic element with post-elastic stiffness given by
equation (2 -2) withfLset to unity, yield displacement determined by the procedures for lead-rubber
isolators and strength Qd approximatelygiven by:
0.0652
eff d M
d d M
K DQ K D
(2 -4)
whereDMis the displacement in the maximum earthquake and eff is the effective damping. Note that the
simplification in equation (2-4), 0.065KdDM, is based on arbitrarily taking the effective damping as 0.04.
Also, the value of Qd may be deduced from the test loops.
2. 3
Modeling Sliding Isolator Units
Spherically shaped sliding isolators come in a variety of configurations, with behavior dictated by the
different combinations of surfaces upon which sliding occurs. The formulation, implementation and
validation of models for multi-spherical sliding isolators can be found in Fenz and Constantinou (2008).
Double and triple Friction PendulumTM
(FP) configurations offer many benefits over the single FP
configuration. For example, more compact isolators, increased displacement capacity, decreased sliding
velocities (approximately halved) and therefore reduced frictional heating and associated problems with
wear and variability in friction. This report focuses on a special type of triple FP isolator (defined in
this section), which is commonly used in practice.
The rigid-linear model (Figure 2-2), is a simple but valuable model that can be used for all types of FP
isolators. The two parameters that govern performance are the characteristic strength at zero displacement
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
26/242
8
Qd and the post-elastic stiffness Kd. Qd is a function of the friction coefficient and the weight W,
whereasKdis a function of the effective radius of curvature of the concave plates Reffand the weight W. It
is an accurate representation for a single FP isolator and for a double FP isolator where the upper and
lower concave plates have the same radius of curvature and same coefficient of friction. For a triple FP
isolator it slightly under-predicts the friction coefficient at zero displacement, however the effect is minor
on the structures global response. Kd is equal to W/(2R1,eff) for double and triple FP isolators when the
upper and lower concave surfaces have an identical radius of curvature.
For this model only the strength needs to be determined from testing, since the stiffness is calculated from
theory. The friction coefficient at zero displacement, , is determined directly from the measured energy
dissipated per cycle (normalized by the weight) divided by the distance travelled. Therefore there is no
interpretation/fitting of test data required. This model was used for the preliminary design of the triple FP
isolation system and also when comparing the nominal properties from production isolators.
Figure 2-2 Rigid-Linear Force-Displacement Behavior of FP Isolators
The rigid-linear model, or use of a bilinear model in Figure 2-1, will give a reasonable estimate of the
global response of the structure. If in-structure accelerations and residual displacement are of interest to
the Registered Design Professional (RDP) then it is appropriate to adopt a more sophisticated model of
the triple FP isolator, such as the tri-linear model described in the following.
The triple FP isolator has a complex behavior with various force-displacement regimes. As the
displacement increases, there are multiple changes in stiffness and strength. Typically at low levels of
force/displacement the system is very stiff, compared to a design level event where there is a lower
stiffness and increased damping, and then different behavior again at the maximum considered event
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
27/242
9
where there is a higher stiffness to control displacements. This gives rise to an adaptive system which can
be optimized for various performance levels.
There are 12 geometric parameters and 4 frictional parameters which define the force-displacement
behavior of the triple friction pendulum isolator, outlined in Figure 2-3. Here the parameters are
numbered from 1 to 4 from the bottom up.
Figure 2-3 Triple Friction Pendulum Cross Section, Definition of Parameters
The three types of geometric parameters are the radius of the concave surface/plate Ri, the height of
various components hiand the distance diwhich is related to the displacement capacity. The stiffness of
friction pendulum isolators is entirely dependent on the weight and some combination of the effective
radius of curvature of the concave plates. The effective radius for each surface is the distance to the pivot
point which, for a double and triple friction pendulum isolator, is always less than the geometric radius
and calculated as:
, , 1, 2, 3, 4i eff i iR R h for i (2 - 5)
This is then used to calculate the actual displacement capacity of each sliding surface, given by:
,* , 1, 2, 3, 4i eff
i i
i
Rd d for i
R (2 - 6)
The characteristics of the special type of friction pendulum isolator, used in this report and commonly
used in practice, are as follows:
R1=R4>>R2=R3
2= 3< 1= 4
d1= d4and d2= d3
h1= h4and h2= h3
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
28/242
10
Essentially the properties of the isolator are a mirror image about the mid-height of the isolator. This
reduces the number of variables from 16 to 6 geometric parameters and 2 frictional parameters. By
adopting these parameters the force-displacement behavior is reduced to three regimes. A description of
these regimes and force-displacement relationships is given in Table 2-1 and 2-2.
Table 2-1 Force-Displacement Behavior for a Special Triple FPR Isolator
Regime Description Force-Displacement Relationship
ISliding on surfaces 2 and 3
only
2
2,2 eff
WF u W
R
Valid for displacement, u:*0 u u
Where: * 1 2 2,2 effu R
II
Motion stopped on surfaces 2
and 3;
Sliding on surfaces 1 and 4
* 1,1
2 eff
WF u u W
R
Valid for displacement, u:* **
u u u
Where:** * *
12u u d
III
Slider bears on
restrainer of surfaces 1 and 4;
Sliding on surfaces 2 and 3
** ** * 12, 1,
( )2 2eff eff
W WF u u u u W
R R
Valid for displacement, u:**
cap
u u u
Where:* *
1 22 2
capu d d
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
29/242
11
Table 2-2 Schematics of Force-Displacement Behavior for Special Triple FP Isolator
Regime Schematic of Isolator Force-Displacement Behavior
I
II
III
Assuming that the isolator does not reach Regime III for MCE displacements, the force-displacement can
be modeled by that shown in Figure 2-4. The force at zero displacement is given by:
2,1 1 21,
eff
eff
RW W
R
(2 - 7)
LateralForce
Displacement
Regime I
LateralFor
ce
Displacement
Regime II
LateralForce
Displacement
Regime III
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
30/242
12
Figure 2-4 Tri-linear Force-Displacement Behavior of Special Triple FP up to Regime II
This model requires two friction coefficients to be determined from testing. As shown in Section 7, these
parameters should be used with care given the uncertainty in predicting them. The stiffness can be
determined based on theory.
Sarlis and Constantinou (2010) present guidelines on how to model triple friction pendulum isolators in
the program SAP2000. For this report the simpler, parallel model is adopted for analysis as it can
accurately capture the behavior as illustrated in Figure 2-4. This model is described in more detail in
Section 9. It is based on the use of two single FP elements. Program SAP2000 also has a new triple FP
element which is not used in this report.
2. 4
Property Modification Factors
The concept of using property modification (lambda) factors to bound the likely isolator response was
originally presented in by Constantinou et al (1999) together with information on the lifetime behavior of
isolators and recommended lambda factors. Shortly after Thompson et al (2000) presented additional
data to better understand the behavior of elastomeric isolators and the related lambda factors. The
concept was first implemented in the AASHTO Guide Specification for Seismic Isolation Design (1999)
and later in the European Standard (2005). It is noted that both of these Standards are for bridges, and that
ASCE 41-2013 and ASCE 7-2016 are the first formal application to existing and new buildings,
respectively. This is because bridge isolators are typically exposed to much more severe environmental
and loading conditions (i.e. cumulative movement) than building isolators, and therefore have a larger
variation from the tested nominal properties. The latest knowledge on lambda factors is now presented in
Constantinou et al (2007), which is a revision of the original 1999 document.
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
31/242
13
The actual mechanical properties of an isolation system during an earthquake will be different to the
nominal values assumed in analysis. Some reasons for this difference may be because:
1) The exact time of an earthquake is unknown and so too is the exact state of the isolator and the
effects of aging, contamination, ambient temperature, and the like.
2) The force-displacement model adopted for analysis may assume that the nominal values remain
constant during seismic motion (as in this report), whereas in reality the properties are changing
instantaneously due to the effects of heating, rate of loading, scragging, etc.
3) The nominal values used for analysis may be based on testing of two prototype isolators, which
will have a different average nominal value to that of all production isolators due to the effects of
manufacturing variation.
Section 3 gives an overview of the effects that have an impact on the isolator properties. To account for
these effects, a rational procedure is to form probable maximum and minimum values of the isolators
mechanical properties which envelope the likely response. One approach is to conduct a statistical
analysis of the distribution of the properties and consideration of the likelihood of occurrence of relevant
events, including the design seismic event. However, a simpler and more practical approach is to assess
the impact of a particular effect (i.e. heating, aging, velocity of loading, vertical load, etc.) on the isolators
properties through testing, rational analysis and engineering judgment (Constantinou et al 2007). If these
effects are appreciable at influencing an isolators force-displacement behavior (i.e. properties Kd and/or
Qd), then it is then quantified in the form of a property modification (lambda, ) factor.
The product of the lambda factors and the nominal properties of the isolator give the upper or lower
bound of what is expected over the isolators design life (including during earthquake excitation). ASCE
7-2016 defines the nominal design properties as the average properties over three cycles of motion and
categorizes the lambda factors into three types (see Section 3), which are:
aewhich accounts for aging and environmental effects.
testwhich accounts for heating, rate of loading, and scragging.
specwhich accounts for permissible manufacturing variations.
The three categories above are then divided into the maximum and minimum values, which may be made
up of one or a series of lambda factors. For the aging and environmental lambda factor, this is shown
indicatively as follows:
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
32/242
14
,max ,1 ,2 ,ae max max max n (2 - 8)
,min ,1 ,2 ,ae min min min n (2 - 9)
Each of the individual values of max,i (i = 1 to n) is greater than or equal to 1.0, whereas each of the
individual values of min,i(i = 1 to n) is less than or equal to 1.0.
A set of six lambda factors ae,max, ae,min, test,max, test,min, spec,maxand spec,minare determined for each
parameter which defines the force-displacement behavior of the isolator. For example, two sets for a lead-
rubber isolator bilinear model are needed: one set of six for the post-elastic stiffnessKdand one set of six
for the characteristic strength Qd. The three max lambda factors are then combined to obtain the
maximum system factor max, whereas the three min lambda factors are combined to obtain the
minimum system factor min. The simple multiplication of the respective three lambda factors might result
in a system factor that is overly conservative. That is, the probability that several additive effects (i.e.
lowest temperature, maximum aging, etc.) occur simultaneously with the maximum considered
earthquake is considered very small. Therefore the system factor should be adjusted based on a statistical
analysis of the variations in mechanical properties with time, the probability of occurrence of joint events
and the significance of the structure. Constantinou et al 1999 proposed a simple procedure for routine
implementation which uses an adjustment factorfawhich is adopted in ASCE 7 as follows:
max ,max ,max ,max(1 ( 1))a ae test specf (2 - 10)
min ,min ,min ,min(1 (1 ))a ae test specf (2 - 11)
The value offais based on the significance of the project and engineering judgment. Values range from
0.66 for a typical structure to 1.0 for a critical facility. ASCE 7-2016 uses a default value of 0.75 which is
only applied to the aging lambda factor, ae. The rationale for use of the adjustment factor is that the full
aging and contamination effects over the lifetime of the structure will likely not be realized when the
structure is subjected to the controlling earthquake event. However, the RDP may choose to adjust the
value, say to 1.0 for significant structures.
The upper and lower bound is then the product of the system lambda factor and its respective nominal
property. For the post-elastic stiffnessKdthat is:
,max max, , mindd K d no al K K (2 - 12)
,min min, , mindd K d no al K K (2 - 13)
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
33/242
15
The same process is followed for the characteristic strength Qd. This process may be applied to other
parameters of the model used in response history analysis when that model is more complex than the
rigid-linear or bilinear hysteretic model. The upper and lower bound force-displacement characteristics
using a bilinear analysis model are illustrated indicatively in Figure 2-5. The effective stiffness is denoted
by KM.
Figure 2-5: Upper and Lower Bound Isolator Force-Displacement Properties
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
34/242
16
This Page is Intentionally Left Blank
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
35/242
17
SECTION 3
TYPES OF PROPERTY MODIFICATIONS
3. 1
Introduction
The properties of seismic isolators will vary over the design life due to the effects of aging and
contamination; will vary during seismic motion due to the effects of heating, history of loading, and
scragging; and will vary between isolator units due to manufacturing differences. Some of these effects
will increase the isolators stiffness and strength (or by relation, the effecting period and effective
damping), whereas others will decrease these parameters. This section presents the procedure to calculate
nominal properties (Section 3.3.1) and gives an overview of how the mechanical properties of low-
damping rubber and lead-rubber elastomeric isolators and Friction PendulumTM
sliding isolators are
influenced by different effects.
The Standard ASCE 7-2016 has categorized the varieties of property modifications (lambda, factors)
into three types: aging and environmental, testingand the specification lambda factors- ae., testand spec,
respectively. The sections that follow are titled according to these three categories of lambda factors,
specifically:
Aging and Environmental Effects- ae.max and ae.min (Section 3.2)
Hysteretic Heating and History of Loading Effects- test,max and test,min (Section 3.3)
Permissible Manufacturing Variation- spec,max and spec,min (Section 3.4)
These lambda factors are determined from a combination of qualification, prototype and production test
data, where the description of each of these testing regimes is given in Section 4.
3. 2
Aging and Environmental Effects: ae,max and ae,min
The aging and environmental lambda factors ae.max and ae.minaccount for the change in properties that
occur over the isolators design life. The effects cannot be quantified by prototype or production testing,
but are developed by a combination of qualification testing, theory and analysis. The aging and
environmental lambda factors account for, but are not limited to, aging, creep, contamination, fatigue,
effects of ambient temperature and cumulative travel. For buildings (and only the types of isolator used in
this report), aging and contamination are the relevant considerations. This is assuming there is little/no
movement in the isolators due to service loads (i.e. wind) and that the isolators are not exposed to extreme
temperatures or damaging substances, which is generally the case. The effects of aging and contamination
are typically always greater than unity (that is ae.max >1.0 andae.min= 1.0). Cumulative travel, fatigue
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
36/242
18
and low temperatures are an issue for bridges and are not considered further. Creep may be an issue for
improperly designed isolators - an issue not addressed in this report. The interested reader may see
Constantinou et al (2007) for a discussion.
3.2.1
Elastomeric Isolator Considerations
Aging in elastomeric isolators depends on the rubber compound and the quality of vulcanization and
curing of the rubber, therefore it is manufacturer specific. Over time the elastomers harden due to
continued vulcanization of the rubber matrix, causing both an increase in the post-elastic stiffnessKdand
in the characteristic strength Qd (i.e. both the effective stiffness and effective damping increase). The
increase is expected to be of the order of 10 to 20% over a 30 year period for low-damping, high shear
modulus rubber compounds (0.5-1.0MPa or 72-145psi) according to Buckle et al (2006). For both natural
rubber and lead-rubber isolators, which have low-damping rubber (i.e. 5% damping), the increase in
strength (or damping) of the rubber due to aging, in proportion to the isolation systems strength (i.e. from
lead), is minor. However the effects of strength may warrant consideration if the isolation system has a
large number of natural rubber isolators compared to lead-rubber isolators.
For lower shear modulus rubbers ( 0.5MPa or 72psi) and for inexperienced manufacturers, the aging
can be significant since low shear modulus rubber can be produced by incomplete curing. This can
perhaps be detected in prototype and/or production testing by observing large scragging effects. This is
because scragging is believed to also be a result of incomplete curing and hence is associated with aging
and continuing chemical processes in the rubber. In the ASCE 7-2016 commentary, a default aging
lambda factor a,max of 1.3 is applied to the post-elastic stiffness of low-damping elastomeric and lead-
rubber isolators over the life of the structure. As discussed, this may not be conservative in some cases.
Conversely, some experienced manufacturers are able to produce low-damping, low shear modulus
rubbers with an aging factor as little as 10%, i.e. a,max=1.1 (Constantinou et al, 2007).
Lead in quality isolators is specified with 99.99% purity and does not experience aging over the design
life of the structure, therefore the aging lambda factors associated with the characteristic strength Qdare
set to unity. Furthermore, contamination is not a concern for elastomeric isolators and the contamination
lambda factors are taken as c,max= c,min=1.0.
Table 3-1 gives the maximum and minimum aging and environmental lambda factors used for the three
design scenarios described in this report (Section 5.4). The lambda factors for Case B and C are based on
the isolators being supplied by a reputable manufacturer. The prototype and production test data is not
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
37/242
19
required to determine the ae factors, however is noted in Table 3-1 for consistency with the design
scenarios/cases listed in Section 5.4.
Table 3-1 Aging and Environmental Lambda Factors for Elastomeric Isolators
Case DescriptionLambda
Factor
Low-Damping
Elastomeric IsolatorLead-Rubber Isolator
Kd Kd Qd
ANo Qualification
Test DataDefault Values
ae,max 1.3 1.3 1.0
ae,min 1.0 1.0 1.0
BQualification Data and
Prototype Test Data Available
ae,max 1.1 1.1 1.0
ae,min 1.0 1.0 1.0
CQualification Data and
Production Test Data Available
ae,max 1.1 1.1 1.0
ae,min 1.0 1.0 1.0
Note: ae,max= a,maxc,maxand ae,min= a,minc,min
Later, when multiplying all the lambda factors it is noted that ae is adjusted by a factor (see equations 2-
10, 2-11) to account for the conservative assumption of having full aging and environmental effects when
the governing earthquake occurs. For example, the adjusted ae,max for theKdof both the elastomeric and
lead-rubber isolators, using an adjustment factor of 0.75, is 1.23for default values and 1.08 for the case
where qualification data are available.
3.2.2
Sliding Isolator Considerations
The aging and environmental lambda factors listed herein are determined for a particular type of sliding
interface. Specifically, a sliding interface with unlubricated highly polished austenitic stainless steel in
contact with PTFE (polytetrafluoroethylene) or similar composite materials, which is sealed and placed in
a normal environment (refer to Constantinou et al, 2007 for details).
The lambda factors for other sliding interfaces (lubricated, bimetallic), installation methods (unsealed)
and environments (severe) will generally be greater, sometimes significantly. The reader is referred toConstantinou et al (2007) for guidance. In ASCE 7-2016 commentary, the default lambda factors are
given for two types of interfaces, unlubricated and lubricated. Bimetallic interfaces are discouraged.
Given that there is considerable uncertainty in the lifetime behavior of generic sliding surfaces, the default
lambda factors result in a large range of property modification values.
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
38/242
20
Aging results in increases in the characteristic strength/coefficient of friction. It is a complicated
phenomenon, but primarily is due to corrosion of the stainless steel surface in properly designed isolators.
Here the maximum aging lambda factor a,max is taken as 1.1 with a,min equal to unity. This is for an
exposure time of 30 years and for Type 304 austenitic stainless steel, where a lower value might be
justified if Type 316 is used since it has a higher resistance to corrosion. The ASCE 7-2016 commentary
specifies a default value for a,maxof 1.3 for unlubricated interfaces.
Contamination also gives an increase in characteristic strength /coefficient of friction. It is complex and
apparently caused by third body effects and abrasion of the stainless steel (see Constantinou et al, 2007).
The maximum contamination lambda factor c,max also depends on whether the concave sliding plate is
positioned so it is facing down (i.e. on the top) or facing up (i.e. on the bottom). For a single friction
pendulum (FP) isolator the contamination lambda factor c,maxis taken as 1.0 if facing down, and 1.1 if
configured to face up. For double and triple FP isolators there are multiple sliding interfaces facing up and
down. Fenz and Constantinou (2006) give an equation to calculate the composite factor when there are
two sliding interfaces. When the radius of curvature and coefficient of friction are equal on the top and
bottom concave surface, the equation reduces to c,max= (c,upper + c,lower)/2 = 1.05. ASCE 7-2016
commentary gives a default c,maxof 1.2 and is not specific on whether the sliding surface is facing up or
down. For all cases a,min is equal to unity.
For the triple FP it is also assumed that the same stainless steel overlay is used for the inner and outer
surfaces (i.e. surfaces 1 and 4, and surfaces 2 and 3, respectively-see Figure 2-3) and therefore the lambda
factors above are applicable for both the inner and outer surfaces friction coefficients. Since the majority
of the motion occurs on the outer surfaces, a different material may be used on the inner surfaces for ease
in manufacturing which could affect the aging and contamination lambda factors.
Table 3-2 gives the maximum aging and environmental lambda factors associated with the isolator
characteristic strength Qd (or by relation, the friction coefficient).The post-elastic stiffness Kdof sliding
isolators is based on the geometry of the isolator and is not effected by aging and environmental factors,
hence the ae,max= ae,min= 1.0. The prototype and production test data is not required to determine the
ae factors, however is noted in Table 3-2 for consistency with the design scenarios in Section 5.4.
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
39/242
21
Table 3-2 Aging and Environmental Lambda Factors for Sliding Isolators with Unlubricated
Stainless Steel-PTFE Interfaces
Case DescriptionLambda
Factor
Single FP Double and Triple
FP (Composite)Face Up Face Down
Qdor Qdor
A
No Qualification
Test Data Available
Default Values
a.max 1.3 1.3 1.3
c.max 1.2 1.2 1.2
ae,max 1.56 1.56 1.56
ae,min 1.0 1.0 1.0
BQualification and Prototype
Test Data Available
a.max 1.1 1.1 1.1
c.max 1.1 1.0 1.05
ae,max 1.21 1.1 1.16
ae,min 1.0 1.0 1.0
C
Qualification and
Production Test Data
Available
a.max 1.1 1.1 1.1
c.max 1.1 1.0 1.05
ae,max 1.21 1.1 1.16
ae,min 1.0 1.0 1.0
Note:For brevity, the minimum lambda factors for aging a,minand contamination c.minare not shown as
they are all equal to 1.0.
As before, when multiplying all the lambda factors it is noted that ae is adjusted by a factor (see
equations 2-10, 2-11) to account for the conservative assumption of having full aging and environmental
effects when the governing earthquake occurs. For example the adjusted ae,max for the triple FP isolator,
using an adjustment factor of 0.75, is 1.42 and 1.12 based on the default and qualification data,
respectively.
3. 3 Hysteretic Heating and History of Loading Effects: test,max and test,min
The testing lambda factors test,max and test,minare specific to the adopted nominal properties and account
for the instantaneous change in properties that occur during an earthquake. The effects are bestdetermined based on dynamic testing at relevant high velocity conditions. If quasi-static testing is
undertaken, then a correction shall be applied for the testing velocity, in which qualification test data may
be used to quantify the difference between quasi-static and true dynamic properties. However, the
correction for dynamics effects may be too complex and with uncertainties and so dynamic testing is
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
40/242
22
preferred. The testing lambda factor accounts for, but is not limited to, heating produced during motion,
velocity and strain effects, scragging of virgin rubber and variation in vertical load.
Before discussing how elastomeric and sliding isolators are affected by hysteretic heating and history of
loading effects, it is important to address the issue of bounding.
3.3.1
Nominal Properties and Bounding
The nominal properties are not used for analysis directly- it is the nominal property multiplied by max or
minfactor which is used for analysis. An important part of factors max and minare the components test,max
and test,minwhich are determined in the testing of prototype isolators. These two factors intend to capture
scragging, and variability and degradation in properties due to the effects of heating and speed of motion.
The bounding is required because the typical models used for analysis (described in Section 2) assume
that the properties are constant, whereas in reality the properties instantaneously change as shown by the
test data in Figure 3-1(a). The change in properties depend on the earthquake and isolation systems
characteristics. For a near-fault, pulse-like earthquake there may be only a small reduction in properties
when the isolation system and/or superstructure experience its peak response. In contrast, for a large
magnitude soft-site scenario there may be a substantial reduction in the isolation system properties when
peak responses occur. Simply adopting the nominal properties for analysis may be a conservative lower
bound (and not conservative for the upper bound) in the former case, whereas in the latter case the
nominal properties may not be conservative for the lower bound. Hence bounding is required through the
use of lambda-test factors to adjust the (somewhat arbitrarily defined) nominal properties to a
representative upper and lower bound.
An interpretation on ASCE 7-2016 is as follows:
1) The nominal properties are defined as the average of three cycles of motion at the maximum
displacement DM.
2) The test,max value is defined as the ratio of the first cycle value of the property to the nominal
value. Thus, effectively the first cycle value is utilized in the upper bound analysis.
3) It is implied that the test,min value is the ratio of the third cycle value of the property to the
nominal value. Thus, effectively the third cycle value is utilized in the lower bound analysis.
The authors of this report believe this to be unduly conservative for most cases and that the
number of cycles considered for determining test,minshould be based on rational analysis with due
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
41/242
23
consideration for the properties of the isolation system, the site conditions and the ground motion
characteristics, including information on proximity to the fault.
Further to bullet point 3) above, ASCE 7-2016 gives test specimen adequacy criteria which may
arbitrarily dictate the selection of the test,max or test,min used for analysis. These requirements are
discussed in Sections 6.5 and 7.4 where they are compared to actual test data from two reputable
manufacturers.
Figure 3-1 illustrates the concept of bounding the hysteretic heating and history of loading effects for a
lead-rubber isolator. The philosophy is the same for Friction Pendulum (FP) isolators. Figure 3-1(a) gives
prototype test data where there are three cycles of motion under relevant high-speed conditions and
amplitude corresponding to the MCE displacement DM. Using the procedures described in Section 6, the
measured properties per cycle can be used to plot a bilinear representation (per Section 2.2) for each
cycle, as shown in Figure 3-1(b). This data can then be used to calculate the nominal properties, defined
as the average among the three cycles, and can be used as a basis to determine the bounds for analysis.
Figure 3-1(b) also shows schematically the upper and lower bounds for analysis (i.e. range in properties
during seismic motion) set as the 1stcycle and 2ndcycle properties, respectively. A rational approach to
determine these bounds is explained in the following.
(a)
Prototype Test Data and Bilinear Model of Nominal Properties
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
42/242
24
(b)
Bilinear Model of Each Cycle of Motion with Representative Bounds
Figure 3-1 Concept of Bounding the Hysteretic Heating and History of Loading Effects
The test,maxvalue intends to capture effects during the first cycle of motion (i.e. the upper bound) which
will certainly be experienced during an earthquake and is thus defined as the ratio of the first cycle value
of the property divided by the nominal value. The test,minvalue is meant to provide a lower bound value
of the property related to the expected number of cycles. Typically the expected number of cycles is
small and is dependent on the properties of the isolation system and the seismic excitation.
Representative results are provided in Warn and Whittaker (2004, 2007) from where the Figures 3-2 (a),
(b) and (c) are adopted. The three figures are for three types of ground motion: (a) near-fault, (b) large-
magnitude, small distance and (c) large magnitude, soft-site.
Figure 3-2 shows the average number of equivalent cycles (or normalized energy dissipated- NED) at the
maximum displacement to have the same energy dissipated as in the actual history of bi-directional
motion determined in response history analysis. It is calculated as follows:
FduNED
EDC (3 - 1)
where EDC is the energy dissipated from one fully reversed cycle at the maximum displacement and F is
the isolator lateral force corresponding to an increment of displacement du. Integration ofF.duover the
duration of motion gives the cumulative energy dissipated by the seismic isolation system as obtained
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
43/242
25
from the nonlinear response history analysis. Quantity Qd/W is the characteristic strength divided by the
weight and quantity Tdis the period calculated on the basis of the post-elastic stiffnessKd:
2dd
WT
K g (3 - 2)
Consider for example the case of the analyzed friction pendulum isolation system in this report. The
nominal values of Qd/Wand Td are 0.055 (weighted average of coefficient B, Tables 7-3 and 7-4) and
4.1sec, respectively. Note that these are the nominal properties, not the lower bound values of Qd/Wand
Tdwhich are stated in the preliminary design (Section 5). Considering that for the site of the structure the
motions will be near-fault or large magnitude, short distance, the expected number of cycles is two (see
Figure 3-2 a) and b)). Therefore at the conclusion of the event, the isolation system is expected to have
undergone the equivalent of two fully reversed cycles at the MCE displacement amplitude, and the
second-cycle properties are appropriate (and somehow conservative-to be demonstrated in the sequel) for
the lower bound. For this case, the test,minshould be defined as equal to the second-cycle value divided by
the nominal value for each property of interest. If the site is such that soft soil conditions prevail, the
expected number of cycles is about three (see Figure 3-2 c)), thus the RDP may choose to set test,minequal
to the third-cycle value divided by the nominal value.
For the analyzed elastomeric isolation system, the nominal values of Qd/Wand Td are 0.10 and 2.6sec,
respectively. Considering again that for the site of the structure the motions will be near-fault or large
magnitude, short distance, the expected number of cycles is again two. In this case the expected number
of cycles is also two for the soft-site case.
(a) Near-Fault Ground Motions (Bin 1) (b) Large Magnitude, Small Distance Motions
(Bin 2M)
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
44/242
26
(c) Large Magnitude Soft-Site Motions (Bin 7)
Note: Bin numbers refer to the set of ground motions used in response history analysis
Figure 3-2 Number of Equivalent Cycles at Maximum Displacement for Various Isolation System
Properties based on Bi-Directional Response History Analysis (Warn and Whittaker 2004)
It is noted that Constantinou et al (2011) recommend adopting the nominal values, which are the average
of three cycles, as the lower bound (i.e. test,minset equal to 1.0). This is considered a reasonable approach
because:
a)
The expected number of cycles for typical isolation systems is about two and therefore the three-
cycle average is representative of the lower bound (i.e. second cycle values). Two cycles appears
to be a reasonable single value for all isolation systems cases based on Figure 3-2 and considering
that typical building isolation systems have a Qd/W of around 0.06 or larger and Tdof 2.5 seconds
or greater.
b) The second-cycle values and nominal values are similar as shown in Figure 3-1(a). From test data
in this report the difference was 9% or less.
c)
Studies by Kalpakidis et al (2010) which show that adopting the three-cycle average value as the
lower bound provides conservative estimates of demands. This last point is explained in summary
below.
To illustrate the conservatism of the bounding analysis when accounting for the properties of lead-rubber
isolators over three cycles of motion at the MCE displacement, Kalpakidis et al (2010) utilized a validated
mathematical model which accounts for the instantaneous heating effects on the lead core strength to
compute the response of seismically isolated structures. They were then compared to results obtained by
-
7/25/2019 Property Modification Factors for Seismic Isolators: Design Guidance for Buildings
45/242
27
using bilinear hysteretic modeling of the isolators and performing bounding analysis with due
consideration to only the heating effects. The upper b